Increasing Genetic Gains in Maize through Breeding for Drought Tolerance

Yoseph Beyene, CIMMYT y.beyene@cgiar.org

Presentation to a senior maize breeders training course on Product Profile Based Breeding for Increased Genetic Gains

Sunbird Capital Hotel, Lilongwe, Malawi, November 11-15, 2019

Outline

• Introduction

• Product concept

• Breeding strategies for drought tolerance

• New tools and technologies to accelerate product

development:

– Use of DH technology

– Introgression of Off-PVP

– Application of molecular markers ( MARS /GS)

Maize in Sub-Saharan Africa • Maize is life for

most of Africa

• Grown on 36 million ha in SSA

• > 208 million farmers depend on maize

• Average yields in SSA are the lowest (<2 t/ha) in the world

– Drought, Low N

– Biotic stresses (MLN, FAW)

0

1

2

3

4

5

6

19

61

19

64

19

67

19

70

19

73

19

76

19

79

19

82

19

85

19

88

19

91

19

94

19

97

20

00

20

03

20

06

20

09

20

12

Yiel

d (

MT/

ha)

World

S. America

SE Asia

Ethiopia

SSA

Angola

CIMMYT’s extensive maize phenotyping network in Africa

#CIMMYTsw18

Plant breeding is cyclic processAssemble your

germplasm

Characterize the genetic variation in

your germplasm

Select appropriate parents for breeding

Generate progeny

Evaluate progeny

Select superior progeny

Repeat evaluation and

selection

Release

The result is incremental gain

Product concept

• White hybrids (SC, 3-way, Double-cross, OPV)

• Yield improvement compared to the best commercial checks:

– Managed drought 45 kg/ha/year

– Low N 45 kg/ha/year

– Optimal 110 kg/ha/year

– Per se female yield 45 kg/ha/year

– MLN 250 kg/ha/year

• General Product Requirements

• Maturity : early, intermediate and late

• Disease resistance: GLS, NLB, MSV and MLN

• Agronomic traits: root lodging, maturity, husk cover

• Grain Texture : flint to flinty-dent

• Color

CIMMYT-GMP Breeding ApproachesAccelerate forward breeding – recycling elite lines from different breeding streams

– DT elite line x DT elite line

– DT elite line x Insect resistant elite line

– DT elite line x NUE elite line

– DT elite line x MLN tolerant line

– DT elite line x Off-PVP line

Integrated application of technologies to enhance genetic gains and breeding efficiency

– DH technology

– MARS and GS in breeding pipelines

– Trial networking system

– Decision support tools (META-R, AGD-R, GEA-R,)

Product profiles for targeted agro-ecologies– Descriptors for lines and single cross

– Male and female yield, Flowering synchronization

Defining drought tolerance

In case of maize (as per CIMMYT), “drought tolerance” is the ability of a genotype to show a substantive and statistically significant yield advantage over the widely grown commercial checks (tested in several drought stress locations), and often under severe drought conditions, including flowering and grain filling period.

Defining Drought Stress

▪ Stage of drought stress: Vegetative; Flowering; Grain filling

▪ Type of drought stress: Mild, Moderate, Severe

▪ Interaction with other factors: Heat; Low soil N

Balance of optimum, managed drought stress locations (on-station trials) and natural stress conditions (on-farm trials)

Effect of moisture stress at different growth stage in maize

The aim of managed drought is not to simulate a farmers field, but create a screen that has high probability of occurrence in farmers’ fields and clearly reveals genotypic differences under stress.

Growth stage Susceptibility of yield to drought

Probability of stress Probability of breeding success

Germination, establishment

High Usually high Low

Pre-flowering Low Depending on the environment

Medium

Flowering High Depending on the environment

High

Post-flowering Medium Usually high Medium

Apply so much stress that yield is reduced by at least 50% (better 65-75 %) of the unstressed screening at that site=> Genotype-by-stress interactions become visible

Breeding for drought tolerance – The challenges

– Genotype differences are often smaller under stress and heritability for GY decreases

– G x E interactions become more pronounced under stress conditions

– Complex, polygenic inheritance

– Stress enhances/magnifies field variability

– Every effort must be made to keep experimental error as low as possible

– How to make progress?

• Early-stage screening

– Fewer replications and locations

– Many genotypes

– More MSE screening (?)

• Later-stage screening

– More replications and locations

– Larger or smaller plots (tradeoff)

– Fewer genotypes

– More TPE screening (?)

• Experimental design selection factors

– Field space and dimensions

– Resources vs # entry

– Soil variability patterns

– Screening environment

– Phenotyping plans

– Power analysis criteria and results

– Statistical advice

• Plot-to-plot competition and alley effects

– Insure genotypes that are phenotypically similar are grown adjacent to each other (i.e., minimize confounding competition effects)

– Consider removal of plants along alleys or decrease alley size and assume uniform impacts

Trait 1

Tra

it 2

Trait 1T

rait 2

Select for several

cyclesSelect for several

cycles

Adapted from Bernardo 2010

Breeding strategy: Tandem selection• One trait selected at a time• Separate thresholds for each trait

Often deployed for traits that affect

adaptation or involve stress

tolerances

Trait 1

Tra

it 2

Culling level

Culling level

Breeding strategy: Independent Culling

• Minimum levels of performance (thresholds) for each trait

• May apply different selection intensities for each trait

Adapted from Bernardo 2010

Trait 1

Tra

it 2

Breeding strategy: Selection Indices

• Involves selecting for several traits simultaneously on the • Basis of a single index value• The index is usually linear function of the different traits, with

each trait weighted by its importance• Selection indices differ in how the weights are obtained

Index selection >

independent culling

levels >Tandem

selection

Indirect selection

• Selection for a particular trait based on phenotype for another trait with some association to the first

• Efficiency differs among trait and population (depends on heritability, correlation)

• Usually less efficient than direct selection but can be easier, quicker or cheaper: Examples:• Canopy temperature used to select for drought stress

tolerance (Furbank and Tester 2011)• Leaf senescence below the ear as a secondary trait for low

N tolerance in maize (Banziger and Lafitte 1997)• Yield component characteristic (e.g.100 kernel weight) used

to select for grain yield

pX sX

IihyG /)( 2 =

G(y) = Response to selection for trait y

h2=Heritability

i = Selection intensity

σ= Variance

I = Generation interval = length of breeding cycle

+

Accelerating genetic gains in plant breeding

Fundamentals→ Experimental designs and control of the plot-to-plot variability→ GxE -- correlation among experimental trials (farmer’s field)→ Population size in multi-environment trials.→ High throughput phenotyping→ Genetic diversity

Years/Cycle→ Genomic selection/MARS→ Doubled haploids → Off season nursery

CIMMYT’s success in breeding stress-resilient maize germplasm has four decades of continuous efforts….Population Description Genetic material

DTP-Y and DTP-W

60% lowland tropical, 20%subtropical, 20% temperate

Derived from 25 putative drought tolerance sources (Tuxpeño Sequia, Latente, Michoacan21, Suwan 1, CIMMYT Pops. 22, 32, 62, 64,and 66); 11% landraces; rest from elite drought tolerant sources, Corn Belt hybrids, germplasm from Thailand and S. Africa

La Posta Sequia (LPS)

Lowland tropical, late-maturing, white dent.

Derived from CIMMYT Population 43 C, SR.

Drought and low N stress tolerant germplasm development via recurrent selection was initiated at CIMMYT in 1975

Objective of managed flowering drought stress

Experiments are conducted during a rain-free period and irrigation regime is designed so that delay silkingand ear abortion

• Leaves are severely rolled at flowering

• Anthesis-silking interval: 3 to 8 days

• Ears per plant: 0.3 to 0.7

• Yields: 1 - 3 t/ha (15-25 % well watered)

Irrigation

Irrigation

Planting Flowering Harvest

Objective of managed post-floweringdrought stress

Experiments are conducted during a rain-free period and irrigation regime is designed so that ….

• Anthesis-silking interval is less than 3 days

• Leaf senescence is accelerated

• Yields: < 50% of yield under well-watered conditions at that site

Irrigation

Planting Flowering Harvest

Useful traits under managed drought stress

• Grain yield => increase

• Ears per plant => increase

• Anthesis-silking interval (ASI) => decrease

• Leaf senescence => decrease

• Leaf rolling => decrease

• Tassel size => decrease

Other secondary traits : Leaf and stem elongation rate, canopy temperature, leaf photo-oxidation, leaf chlorophyll concentration, seedling survival under drought, osmotic adjustment, leaf erectness.

Commercial Check (SC513)

CZH132018

Murewa, Zimbabwe

How did DT Maize perform during the recent El Niño?

Slide from Peter Setimela

CIMMYT –derived Hybrids vs. Commercial Hybrids in ESA- Head-to-Head Comparison

• Best 10-15 hybrids from different maturity groups (extra-early, early, intermediate, late) vs. corresponding checks from each of DTMA, WEMA and IMAS nodes in ESA

• 7073 comparisons across different maturity groups.

• Results of combined analyses across Optimum-moisture & N-application; Managed-drought; & Low-N Environments.

• Identified best products cutting across environments vis-à-vis popular commercial hybrids

Development of multiple stress tolerant lines and hybrids

• # lines: 7

• # tester : 7

• Design: Line by tester

• Locations: 7 optimum, 2 drought and 2 Low N sites

Development of multiple stress tolerant lines and hybrids

Entry

GY

(t/ha) Entry

GY

(t/ha) Entry

GY

(t/ha)

Opt DS LN

L6/T4 8.8 L4/T4 5.0 L6/T1 3.2

L2/T1 8.8 L4/T2 4.7 L2/T5 2.9

L6/T1 8.7 L7/T1 4.6 L4/T2 2.9

L6/T6 8.6 L4/T1 4.6 L5/T3 2.9

L2/T4 8.5 L5/T4 4.5 L6/T6 2.8

L4/T1 8.5 L3/T2 4.2 L4/T3 2.7

L6/T5 8.5 L7/T5 4.2 L6/T2 2.7

L2/T6 8.4 L6/T3 4.2 L2/T7 2.6

L6/T7 8.4 L2/T7 4.1 L4/T1 2.6

L4/T6 8.3 L2/T4 4.1 L6/T3 2.6

L2/T5 8.2 L5/T7 4.1 L2/T6 2.6

L4/T2 8.2 L4/T6 4.1 L4/T4 2.6

Mean of

Check

hybrids

5.9 2.8 1.7

LSD 0.97 1.54 0.95

H 0.88 0.3 0.19

EntryGY GY GYOpt DS LN

Top ten hybridsover checks

46% 59% 61%

Top ten over thebest check

5% 10% 31%

The best hybridover checks

50% 79% 82%

Yield improvements

Early maturity : Top 10 hybrids under optimum condition

25

No EntryYield (t/ha)

% increase over the best check

AD EA MOI PH RL SL

Top Ten Entries

1 CKDHH0945 7.7 44.5 64.0 2.3 19.5 241.1 5.1 7.4

2 CKDHH0968 7.6 42.8 64.1 2.1 19.4 242.3 2.9 4.7

3 CKDHH0961 7.5 42.0 64.3 2.5 20.4 228.6 6.1 9.3

4 CKDHH1055 7.5 41.5 62.5 2.1 19.3 223.5 2.9 2.9

5 CKDHH1048 7.5 41.4 62.3 2.4 18.8 228.2 2.8 4.5

6 CKDHH1051 7.4 39.9 61.9 2.1 18.7 224.4 5.7 5.6

7 CKDHH0993 7.4 39.7 63.7 2.5 20.1 221.1 3.4 6.3

8 CKDHH0953 7.4 39.2 63.4 2.4 19.5 229.9 5.5 5.1

9 CKDHH0989 7.3 37.1 63.2 2.4 19.5 220.8 3.1 7.2

10 CKDHH1059 7.2 36.4 62.5 2.1 18.6 221.2 2.7 5.0

Checks

1 DH04 5.3 (35) 62.2 2.6 18.8 227.6 6.0 6.0

2 Duma 43 5.3 (34) 61.8 2.8 17.7 226.5 5.3 4.0

3 DK8031 4.4 (36) 63.0 2.9 18.5 222.4 4.2 4.9

Statistics

nreps 2 2 2 2 2 2 2nlocs 19 18 19 19 17 10 10Gmean 6.9 63.1 2.3 19.3 228.4 4.0 6.5LSD 0.6 1.0 0.4 0.9 7.5 4.0 4.9CV 4.7 0.8 8.1 2.3 1.7 51.2 38.1Heritability 0.9 0.7 0.5 0.7 0.8 0.1 0.1

Early maturity: Top 10 hybrids under managed drought

26

Entry Pedigree Yield % IOBC EA MOI PH RL SL

27 CKDHH1051 4.1 40.6 2.3 18.3 163.8 3.0 3.3

25 CKDHH1049 4.0 36.2 2.0 18.0 162.3 5.0 6.428 CKDHH1053 3.9 34.3 2.0 18.4 159.0 4.0 4.3

4 CKDHH0949 3.9 33.6 3.0 19.5 164.0 6.9 3.831 CKDHH1059 3.7 26.5 2.1 18.4 161.7 3.9 4.124 CKDHH1048 3.7 26.1 2.3 17.8 163.3 3.9 6.21 CKDHH0944 3.6 22.6 3.1 18.6 158.5 3.7 5.9

16 CKDHH0989 3.6 21.4 2.5 18.9 153.0 4.6 8.1

8 CKDHH0961 3.6 21.2 3.2 20.5 161.8 2.6 5.07 CKDHH0955 3.5 20.5 3.0 18.8 158.9 7.2 4.4

32 DH04 2.9 2.6 17.6 159.5 3.4 5.333 Duma 43 2.5 2.7 17.9 157.6 8.2 4.634 DK8031 2.3 3.5 17.5 157.0 5.2 6.0

nreps 2 2 2 2 2 2

nlocs 10 7 9 6 5 5

Gmean 3.4 2.6 18.5 160.1 5.4 6.0LSD 0.6 0.6 1.2 9.5 6.3 5.6

CV 9.7 10.8 3.3 3.0 59.0 47.0Heritability 0.8 0.7 0.6 0.5 0.2 0.1

Use of wide testing network to identify high yielding and stable hybrids

• The medium and early sets were evaluated in Kenya, Tanzania and Uganda in 2016

– Random stress: 5 locations

– Optimum: 29-30 locations

– Analysis done: Single site and combined analysis

– Only combined analysis across three countries is presented, but we have also done analysis within each country to identify hybrids adapted to a specific country

Results of Medium-maturity Evaluated at 29 Optimum Locations in Eastern Africa in 2016

Entry Name

Grain Yield (t/ha)

% Increase over Mean of 7 Checks

50% Anthesis(days)

Root Lodging (#)

Stem Lodging (#)

Bad Husk Cover (%)

Ear Aspect (1-5)

TLB (1-5)

Grain Mois(%)

Plant Height (cm)

GLS (1-5)

Female SC Yield (t/ha)

Nicking Difference (days)

46 CKH160066 7.90 27.8 72.6 2.7 5.4 2.6 2.0 1.8 20.6 229.0 1.4 6.36 -4.88

30 CKH160050 7.68 24.2 69.8 2.7 1.5 1.9 2.3 1.8 19.4 224.5 1.5 8.70 -0.59

37 CKH160057 7.65 23.7 69.6 4.6 2.3 4.3 2.4 1.7 20.8 247.6 1.2 8.03 2.38

36 CKH160056 7.61 23.1 70.2 1.2 3.5 3.4 2.6 1.7 20.5 241.9 1.3 6.36 1.99

41 CKH160061 7.56 22.2 70.2 2.9 2.7 5.7 2.1 1.8 20.3 228.3 1.5 9.27 -0.13

45 CKH160065 7.49 21.2 72.9 6.5 5.0 1.7 1.9 1.7 20.0 227.6 1.5 8.03 -4.27

34 CKH160054 7.46 20.7 69.8 2.9 1.2 2.6 2.2 1.7 20.5 222.2 1.2 6.36 3.06

38 CKH160058 7.41 19.9 70.3 1.7 2.8 7.6 2.3 1.8 20.0 236.8 1.4 7.25 2.51

42 CKH160062 7.36 19.1 71.4 2.4 3.5 7.2 2.5 1.9 20.5 219.2 1.3 7.18 -6.57

23 CKH160043 7.31 18.2 69.5 0.4 1.3 17.9 2.4 1.8 19.3 222.4 1.5 8.03 2.66

54 WE1101 5.85 69.3 1.1 2.1 8.9 2.0 1.9 19.4 217.7 1.6

55 CKH12600 6.25 68.3 1.0 2.2 4.0 2.2 2.2 18.6 212.6 2.6

56 CKH10717 6.71 72.4 1.1 5.2 4.1 2.3 1.8 20.1 223.5 1.8

57 H516 5.69 70.5 1.9 8.9 7.6 2.5 2.0 18.9 236.8 2.1

58 WH505 7.18 73.4 2.3 2.4 2.0 2.0 1.7 20.3 241.4 1.6

59 Pioneer 30G19 5.98 70.1 5.3 6.3 2.5 2.7 2.1 18.2 227.5 2.5

60 Local Check 5.60 72.4 3.8 6.4 4.5 2.3 1.9 18.7 223.2 1.7

Heritability 0.90 0.97 0.70 0.84 0.89 0.88 0.72 0.89 0.97 0.89

Grand Mean 6.73 69.8 2.5 3.3 9.2 2.3 1.9 19.3 218.9 1.7

LSD 0.438 0.66 2.99 2.31 3.68 0.22 0.19 0.63 5.67 0.24

CV 3.3 0.5 60.5 35.4 20.5 4.7 5.1 1.7 1.3 6.9

n Replicates 2 2 2 2 2 2 2 2 2 2

n Locations 29 25 9 18 17 19 11 23 22 14

Use of DH to accelerate genetic gain

• Speed– FAST and FURIOUS!

– Greatly reduces the length of a breeding cycle to create genetically pure lines

– 100% homozygosity without genetic segregation or

– Reduces the frequency of unfavorable recessive alleles (i.e. genetic load)

– Helped us to combine different traits into a single lines within 2-3 generation compared to traditional pedigree breeding which request 7-8 generation of selfing

• Increased response to early selection

– DH lines have higher genetic variance in per se & testcross trials compared to F2 or selfed families (F3 or F4)

– Better estimates of QTL effects with homozygous lines

DH technology can accelerate breeding for drought tolerance

• >92,000 DH lines developed from CIMMYT bi-parental crosses through different projects

• The DH lines were evaluated for per se, reaction to major leaf diseases in Africa

• Testcross were developed and evaluated in stages 1, 2 & 3 and regional trials across locations in different countries

• Hybrids were released in Kenya, Uganda, Tanzania, South Africa and nominated in Mozambique

Large Variability for agronomic traits Anthesis date: 44 - 68 daysSilking date: 43 - 72 daysPlant height : 59 - 207 cmEar height : 25 - 122 cm

Genetic variability for drought tolerance among the DH lines

• Date of planting: 6 June 2012

• Date of last irrigation: 26 July 2012

• Date of harvesting: 10 Oct 2012

Date of planting: 25 May 2014Date of last irrigation:19 July 2014Date of harvesting: 28 Oct 2014

Performance of top 10 yielding lines(DH and pedigree) evaluated across at six locations in East Africa in 2014.

• # inbred lines= 124 (34 are CML and 90 are elite lines)

• Inbred lines were categorized into five era of release

• Year of developed – 1996 to 2013 developed through pedigree and DH technology

• Pre-1996 =12 lines• 1996-2000 =6 lines • 2001-2005 = 27 lines • 2006-2010 = 34 line• 2011-2013 = 45 lines

• Evaluated in 6 sites in Kenya and Uganda

Entry Year of release

Yield(t/h) Across

CKDHL0165 5 4.9

CKDHL0186 5 4.5

CML543 4 4.4

CKDHL0590 5 4.4

CKDHL0282 5 4.4

CML395 2 4.2

CKDHL0364 5 4.2

CKDHL0323 5 4.2

CKDHL0295 5 4.0

CKL08002 4 4.0

Mean 3.0

LSD(0.05) 1.3

Repeatability 0.65

Regasa et.al (2016):Euphytica

DT and very productive doubled haploid line (CKDHL0323): Kiboko Demo, Sep 2017

Five elite DH lines released as CMLs in 2015

34

CML566 CML570

Drought tolerant and susceptible hybrids under managed drought at Kiboko in 2017

Commercial check Released hybrid

Drought tolerant and susceptible hybrids under managed drought at Kiboko in 2018

Introgression of Off-PVP US Temperate lines into CIMMYT’s Tropical Maize Germplasm

OFF-PVP inbred lines BC1-S5 temp introgressed tropical inbred lines

273 Of-PVP lines used• Crossed with selected tropical adapted lines• Evaluated in Stage I, II, III, RT• Hybrids released

Comparison of Off-PVP introgressed and original tropical lines

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

Yie

ld t

/ha

• 5 tropical

adapted lines

were crossed

with ex-PVP

lines

• The original

lines together

with BC1 fixed

lines crossed

with 5 testers

• The hybrids

were evaluated

across seven

optimum sites

• # Locations= 7

• Heritability= 0.95

Comparison of hybrid performance involving temperate-introgressed (ex-PVP) vs. tropical lines

Farmer identified WE6105 (male parent had 25% ex-pvp) as best hybrid in her demo field at Cherangany, Kitale, 3 Oct 2018

What about Marker assisted selection

Untested

Stage 1

Stage 2

Stage 2

Stage 3

Conventional

Plant

Breeding

Molecular

Plant

Breeding

GS/MARS at CIMMYT-Africa maize breeding

• Objective: To improve drought tolerance using GS and MARS

• 34 bi-parental populations (total =6252 F2:3, each with ca. 184 progenies)

• Each pop phenotyped in 2-4 managed water-stressed, 3-4 well-watered environments, and genotyped with 190-286 SNPs

• Genetic gain studies completed for 18 populations• 10 populations using MARS

• Selection at C0 based on phenotype and marker data, followed by one recombination and two selfing

• 8 populations using GS • Selection at C0 based on phenotype only, followed by two

cycles of marker only recombination

C1S1

C1S2

Conduct genetic gain studies (cross C0, C1, C1S1, C1S2 , F5:6 and founder

parents with a single-cross tester and evaluate the testcrosses and commercial

checks in multi-location water-stress and well-watered environments

F3:4

Parent 1 x Parent 2 for each population

F1

F2

F2:3 = C0Make tester

cross (TC)

QC genotyping

Polymorphism screening

Genotype C0

Self selected plants

Self selected plants

Self selected plants

C1

Develop Lande-Thompson selection index and select the

best 8 C0 families for next cycle

Select the best individuals

using markers

Evaluate TC under WS and

WW environments

F4:5

F5:6

Select the best individuals

using markers

Select the best individuals

using markers

MARS/GS work flow

MARS and GWS Recombination nurseries

Leaf sampling Pollination

Critical steps:

1. Planning and coordination

2. Data turnaround time from leaf sampling to

genotyping)

3. Data analysis and selection

4. Synchrony and pollination

Gain in grain yield using genomewide SNPs under drought environments in SSA

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Gra

in y

ield

(M

g/h

a)

Populations

Cycle0 Cycle1 Cycle2 Cycle3

Genetic gain from conventional breeding in Africa 18 kg ha-1 yr-1 (Edmeades, 2013)32 kg ha-1 yr-1 (Masuka et al , 2017)

Genetic gain from GS is 2-4 times higher than the gains from conventional breeding reported in sub-Saharan Africa.

Overall gain in GY : 70.5 kg ha-1 year-1

Performance of hybrids developed from GS-C3-DH lines, founder parents, and commercial checks under optimum

Yiel

d; t

/ha

Populations

0

1

2

3

4

5

6

7

8

9

Pop1-GWS Pop2-GWS Pop3-GWS Pop4-GWS Pop5-GWS

Top 5 hybrids Parents Commercial checks

Phenotyped under optimum locationsGain over the commercial checks= 8.7 to 20.8%Gain over the parents = 9.0 to 91.1%

Performance of hybrids developed from GS-C3-DH lines, founder parents and commercial checks under drought

Yiel

d; t

/ha

Populations

• Phenotyped under managed drought locations• Gain over the commercial checks= 47.3 to 98.2%• Gain over the parents = 17.2 to 57.3%

0

0.5

1

1.5

2

2.5

Pop1-GWS Pop2-GWS Pop3-GWS Pop4-GWS Pop5-GWS

Top 5 hybrids Parents Commercial checks

Overall gain in grain yield under drought and optimum across 10 MARS pops

Genetic gain from conventional breeding in Africa 18 kg ha-1 yr-1 (Edmeades, 2013)32 kg ha-1 yr-1 ( Masuka et al 2017)

Genetic gains from MARS are 1.6- 2.8 times higher than from conventional breeding reported in sub-Saharan

Africa.

Overall gain in GY : 105 kg ha-1 year-1 under optimum and 51 kg ha-1 year-1

Performance of hybrids developed from individual C1S2 of MARS

• For each of the 10 populations,

• 47-74 C1S2 lines were extracted after 3 cycles of MARS

• The best five S5 lines developed through pedigree selection

• The founder parents for each population

• The above lines were crossed with a single-cross tester from the opposite heterotic group.

• Five commercial checks

• All the hybrids evaluated under managed stress and optimum conditions with two replications

Performance of hybrids developed from individual C1S2 of MARS and pedigree breeding under drought

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

1008 1015 1016 1017 1018 1019 1020 1021 1023 1028

Gra

in y

ield

(t/

ha)

Populations

Best C1S2

Pedigree

Check

Parents

Beyene et al. (2016) Euphytica : 208:285–297

Heritability = 0.32 to 0.75

Performance of hybrids developed from individual C1S2 of MARS and pedigree breeding

under optimum

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

1008 1015 1016 1017 1018 1019 1020 1021 1023 1028

Gra

in y

ield

(t/

ha)

Populations

Best C1S2

Pedigree

Check

Parents

Heritability = 0.10 to 0.49

Beyene et al. (2016) Euphytica : 208:285–297

Implementation of GS in stage I trials

• # lines genotyped (Stage I)= 1492

• # lines phenotyped = 855

• # sites= 3 optimum and one managed drought

• The phenotypic data were used to predict the remaining untested lines.

R² = 0.7249

4

4.5

5

5.5

6

6.5

7

7.5

8

8.5

3 4 5 6 7 8 9 10

Pre

dic

ted

Observed

Obs vs Pred -Drought

Obs vs Pred Linear (Obs vs Pred)R² = 0.7582

1.5

2

2.5

3

3.5

4

4.5

5

0.5 1.5 2.5 3.5 4.5 5.5

Pre

dic

ted

Observed

Obs vs Pred-Optimum

Obs vs Pred Linear (Obs vs Pred)

Comparison of hybrids advanced based of PS and

GS in stage II trialsCategory # lines # testers # of hybrids

All stage II hybrids 347 3 1042

Hybrids advance through phenotype 175 3 526

Hybrids advance through GEBV 172 3 516

Beyene et al (2019). Frontiers in

Plant Science

GS reduced the cost by 32% over PS

with similar selection gains

Genetic Gain: Intermediate maturity maize across optimum and drought Locations (2008-2017)

# of locations : 37 optimum, 7 drought locations (in Ken, Tan and Uga in 2017 and 2018)

Country Gain kg/ha/year Reference

Argentina 132 Luque et al., 2006

Eastern and southern Africa 109.4 Masuka et al. 2017a

China 94.7 Ci et al., 2011

Canada 80 Bruulsema et al., 2000

United States 65-75 Duvick, 2005

West Africa 40Badu-Apraku et al. (2013,

2015)

This study 131 Unpublished

MAIZE Varietal Releases in 2018

• 81 unique varieties released across Africa, Asia and LatAm in 2018 (63 varieties based on CIMMYT germplasm; 18 based on IITA germplasm)

• 14 varieties are Combination Hybrids (majority from SSA)• 20 of the released varieties are nutritionallly enriched (ProA/QPM/QPM+High Zn)

3.5 million smallholder farmers planted stress tolerant MAIZE varieties in 10 target countries in

Africa (2018)

373617

951,092

267,883

740,789

539,880

84,000

205,760

40,000 94,488

218,341

0

100000

200000

300000

400000

500000

600000

700000

800000

900000

1000000

Kenya Ethiopia Tanzania Uganda Nigeria Ghana Mali Benin Mozambique Zambia

A Breeding program’s hierarchy of needs

Rutkoski, 2018

Thank you

for your

interest!